[v2] powerpc/pci: Assign fixed PHB number based on device-tree properties

The domain/PHB field of PCI addresses has its value obtained from a
global variable, incremented each time a new domain (represented by
struct pci_controller) is added on the system. The domain addition
process happens during boot or due to PCI device hotplug.

As recent kernels are using predictable naming for network interfaces,
the network stack is more tied to PCI naming. This can be a problem in
hotplug scenarios, because PCI addresses will change if devices are
removed and then re-added. This situation seems unusual, but it can
happen if a user wants to replace a NIC without rebooting the machine,
for example.

This patch changes the way PCI domain values are generated: now, we use
device-tree properties to assign fixed PHB numbers to PCI addresses
when available (meaning pSeries and PowerNV cases). We also use a bitmap
to allow dynamic PHB numbering when device-tree properties are not
used. This bitmap keeps track of used PHB numbers and if a PHB is
released (by hotplug operations for example), it allows the reuse of
this PHB number, avoiding PCI address to change in case of device remove
and re-add soon after. No functional changes were introduced.

Signed-off-by: Guilherme G. Piccoli <gpiccoli@linux.vnet.ibm.com>
---
 arch/powerpc/kernel/pci-common.c | 43 +++++++++++++++++++++++++++++++++++++---
 1 file changed, 40 insertions(+), 3 deletions(-)

On Tue, Mar 15, 2016 at 07:54:41PM -0300, Guilherme G. Piccoli wrote:
>The domain/PHB field of PCI addresses has its value obtained from a
>global variable, incremented each time a new domain (represented by
>struct pci_controller) is added on the system. The domain addition
>process happens during boot or due to PCI device hotplug.
>
>As recent kernels are using predictable naming for network interfaces,
>the network stack is more tied to PCI naming. This can be a problem in
>hotplug scenarios, because PCI addresses will change if devices are
>removed and then re-added. This situation seems unusual, but it can
>happen if a user wants to replace a NIC without rebooting the machine,
>for example.
>
>This patch changes the way PCI domain values are generated: now, we use
>device-tree properties to assign fixed PHB numbers to PCI addresses
>when available (meaning pSeries and PowerNV cases). We also use a bitmap
>to allow dynamic PHB numbering when device-tree properties are not
>used. This bitmap keeps track of used PHB numbers and if a PHB is
>released (by hotplug operations for example), it allows the reuse of
>this PHB number, avoiding PCI address to change in case of device remove
>and re-add soon after. No functional changes were introduced.
>
>Signed-off-by: Guilherme G. Piccoli <gpiccoli@linux.vnet.ibm.com>
>---
> arch/powerpc/kernel/pci-common.c | 43 +++++++++++++++++++++++++++++++++++++---
> 1 file changed, 40 insertions(+), 3 deletions(-)
>
>diff --git a/arch/powerpc/kernel/pci-common.c b/arch/powerpc/kernel/pci-common.c
>index 0f7a60f..d4217f9 100644
>--- a/arch/powerpc/kernel/pci-common.c
>+++ b/arch/powerpc/kernel/pci-common.c
>@@ -44,8 +44,11 @@
> static DEFINE_SPINLOCK(hose_spinlock);
> LIST_HEAD(hose_list);
>
>-/* XXX kill that some day ... */
>-static int global_phb_number;		/* Global phb counter */
>+/* For dynamic PHB numbering on get_phb_number(): max number of PHBs. */
>+#define	MAX_PHBS	8192
>+
>+/* For dynamic PHB numbering: used/free PHBs tracking bitmap. */
>+DECLARE_BITMAP(global_phb_bitmap, MAX_PHBS);
>

I guess it can be file-scoped variable.

> /* ISA Memory physical address */
> resource_size_t isa_mem_base;
>@@ -64,6 +67,39 @@ struct dma_map_ops *get_pci_dma_ops(void)
> }
> EXPORT_SYMBOL(get_pci_dma_ops);
>
>+static int get_phb_number(struct device_node *dn)
>+{
>+	const __be64 *prop64;
>+	const __be32 *regs;
>+	int phb_id = 0;
>+
>+	/* try fixed PHB numbering first */
>+	if (machine_is(pseries)) {
>+		regs = of_get_property(dn, "reg", NULL);
>+		if (regs)
>+			return (int)(be32_to_cpu(regs[1]) & 0xFFFF);
>+	}
>+
>+	else { /* maybe PowerNV? */
>+		prop64 = of_get_property(dn, "ibm,opal-phbid", NULL);
>+		if (prop64)
>+			return (int)(be64_to_cpup(prop64) & 0xFFFF);
>+	}
>+

We usually have something like below. Also, why not checking if it's
PowerNV platform by machine_is(powernv)? Besides, I guess the comments
here can be more indicative:

	if (...) {

	} else ( ...) {

	}

>+	/* dynamic PHB numbering mechanism */
>+	while ((phb_id < MAX_PHBS) && test_bit(phb_id, global_phb_bitmap))
>+		phb_id++;
>+
>+	BUG_ON(phb_id >= MAX_PHBS); /* reached maximum number of PHBs */
>+	set_bit(phb_id, global_phb_bitmap);
>+	return phb_id;

It could be simplified to:

        do {
		phb_id = find_next_zero_bit(global_phb_bitmap, MAX_PHBS, 0);
                BUG_ON(phb_id >= MAX_PHBS);
        } while(test_and_set_bit(phb_id, global_phb_bitmap));

	return phb_id;
>+}
>+
>+static void release_phb_number(int phb_id)
>+{
>+	clear_bit(phb_id, global_phb_bitmap);
>+}
>+

The logic of this function can be merged to pcibos_free_controller() as it's
just one statement and it's called by pcibios_free_controller() only. Also,
it will cause memory corruption if (phb_id >= MAX_PHBS) on pSeries/PowerNV
though it's rare.

> struct pci_controller *pcibios_alloc_controller(struct device_node *dev)
> {
> 	struct pci_controller *phb;
>@@ -72,7 +108,7 @@ struct pci_controller *pcibios_alloc_controller(struct device_node *dev)
> 	if (phb == NULL)
> 		return NULL;
> 	spin_lock(&hose_spinlock);
>-	phb->global_number = global_phb_number++;
>+	phb->global_number = get_phb_number(dev);
> 	list_add_tail(&phb->list_node, &hose_list);
> 	spin_unlock(&hose_spinlock);
> 	phb->dn = dev;
>@@ -94,6 +130,7 @@ EXPORT_SYMBOL_GPL(pcibios_alloc_controller);
> void pcibios_free_controller(struct pci_controller *phb)
> {
> 	spin_lock(&hose_spinlock);
>+	release_phb_number(phb->global_number);
> 	list_del(&phb->list_node);
> 	spin_unlock(&hose_spinlock);
>

Thanks,
Gavin

Thanks very much for the comments Gavin! I agree with all of them, will 
rework the patch and send a new version.

Cheers,


Guilherme